Water-wind-light short-term complementary scheduling method under finite probability distribution information

Abstract

The invention discloses a water-wind-light short-term complementary scheduling method under finite probability distribution information, and relates to the field of short-term scheduling of a novel power system. Collecting and sorting initial calculation data, and constructing a moment fuzzy set of new energy output prediction errors by using finite probability distribution information; according to the moment fuzzy set, constructing a hydroelectric flexibility reserve constraint condition based on distribution robust chance constraint planning, and quantifying the hydroelectric reserve capacity; taking the minimum average absolute value of the average distance of the residual load of the power grid as a target function, and providing flexible reserve to compensate a new energy output prediction error while peak regulation is carried out on cascade hydropower; introducing a Chebyshev inequality to convert the distribution robust opportunity constraint to obtain a deterministic linear programming model; and solving the model by using an optimization solver. According to the method, the peak regulation flexibility and the standby flexibility can be provided by reasonably utilizing the cascade hydropower, so that the problem of reverse peak regulation caused by large-scale new energy grid connection and the problem of power imbalance caused by new energy output prediction errors are effectively solved.

Description

Technical Field

[0001] The present invention belongs to the field of short-term dispatching of new power systems, and relates to a short-term complementary dispatching method for water, wind and solar power under finite probability distribution information, and in particular to a short-term complementary dispatching method that fully utilizes the flexibility of cascade hydropower to cope with the uncertainty of wind and solar power output with finite probability information. Background Art

[0002] As the penetration rate of renewable energy in the power system continues to increase, the uncertainty of the power system also increases. In order to ensure the safe and stable operation of the power system, improving the flexibility of the power system to cope with the uncertainty caused by large-scale renewable energy grid connection has become an urgent problem to be solved.

[0003] Renewable energy represented by wind power and photovoltaics has random characteristics such as volatility and intermittency, which brings challenges to the short-term dispatch of new power systems. On the one hand, the output of new energy has anti-peak characteristics, and may fluctuate violently in a short period of time, resulting in drastic changes in the net load curve and increased peak-to-valley difference, which poses a challenge to peak regulation. On the other hand, due to the existence of prediction errors, there is a deviation between the predicted output of wind and solar power and the actual output, which ultimately leads to power imbalance and affects the safe operation of the system. In terms of uncertainty optimization modeling, due to factors such as climate change, the sample data of wind and solar power are constantly changing, so the time span of the sample data used for modeling should not be too long. However, it is difficult to accurately estimate the probability distribution using less sample data, which increases the difficulty of wind and solar uncertainty modeling.

[0004] In view of the above problems, the present invention proposes a short-term complementary scheduling method for water, wind and solar power under finite probability distribution information. This method uses the finite probability distribution information of the new energy output (moment information of the sample) to model its uncertainty, and considers that the probability distribution of the prediction error of the new energy output is unknown and is located in a moment fuzzy set. In the short-term complementary scheduling model of water, wind and solar power, with peak load regulation as the objective function, the process that the hydropower flexibility reserve should compensate for the wind and solar power output prediction error is constructed as a distributed robust opportunity constraint based on the moment fuzzy set, so as to achieve cascade hydropower to provide peak load flexibility and reserve flexibility at the same time. The Chebyshev inequality is introduced to convert the distributed robust opportunity constraint into a definite linear programming constraint. Finally, the water, wind and solar multi-energy complementary system in a certain province is used as the engineering background to verify the effectiveness of this method and its good application effect. Summary of the invention

[0005] The purpose of the present invention is to provide a short-term complementary scheduling method for water, wind and solar power under finite probability distribution information. When giving priority to the consumption of new energy, cascade hydropower is used for peak regulation, and the hydropower flexibility reserve constraint (distributed robust opportunity constraint) is constructed based on the finite probability distribution information of the mean and variance (first-order moment and second-order moment) of the wind and solar power output forecast error. In order to obtain an optimization model that is easy to calculate, based on the Chebyshev inequality, the hydropower flexibility reserve constraint is converted into a deterministic linear constraint condition, and the hydropower flexibility reserve capacity that needs to be configured in each time period is also determined.

[0006] The technical solution of the present invention:

[0007] A short-term complementary scheduling method for hydropower, wind power and solar power under finite probability distribution information, the steps are as follows:

[0008] Step 1: Collect initial calculation conditions, including basic operating parameters of hydropower stations, coupled dispatching demand conditions of power system and cascade hydropower stations, forecast data of photovoltaic output, wind power output, load and interval water inflow at 96 points in the future day, and 96 points of measured historical data and forecast data of wind power output and photovoltaic output in the past three years.

[0009] Step 2: Analyze and organize the measured historical data and forecast data of wind power output and photovoltaic output in the past three years, and calculate the relative forecast error data of wind power output and photovoltaic output in the past three years. These relative forecast error data constitute a sample data set of uncertainty of new energy output, and the sample data set is divided into in-sample data and out-of-sample data according to the proportion. The output forecast error of new energy is the main source of uncertainty, and the flexibility demand of the power grid will be related to the forecast error of new energy output.

[0010] Step 3: Aggregate all wind farms and photovoltaic power stations connected to the grid into a virtual wind farm and a virtual photovoltaic power plant respectively. Utilize cascade hydropower to provide peak load flexibility and reserve flexibility, and couple cascade hydropower with wind and solar power to achieve short-term complementary scheduling. With the goal of minimizing the absolute value of the average deviation of residual load, a short-term complementary scheduling model for hydropower, wind and solar power is constructed:

[0011]

[0012] Where: F is the objective function value, MW; C t , and are the original load, net load and residual load of the power grid in period t, MW; P t AW and P t AS are the predicted wind power output and photovoltaic power output in period t, MW; is the output of hydropower station m in period t, MW; C R,aveis the average value of residual load, MW; T and M are the total number of time periods and the total number of hydropower stations respectively. Introducing auxiliary variable δ t , perform equivalent transformation on formula (1):

[0013]

[0014] The constraints are as follows:

[0015] 1) Water balance constraints

[0016]

[0017] R m,t =q m,t +s m,t (9)

[0018] Where V m,t is the storage capacity of hydropower station m at the end of period t, m 3 ;I m,t and R m,t are the interval water flow and outflow flow of hydropower station m in period t, respectively. 3 / s; where the outbound flow includes the power generation flow q m,t and the abandoned water flow s m,t , m 3 / s; Δt represents the step length of the scheduling calculation period, which is 15 minutes; U m is the index set of the upstream power stations of power station m, j is the upstream power station number, R j,t is the outflow of upstream power station j in period t, m 3 / s. Since water abandonment is not allowed in hydropower stations, the abandonment flow rate is set to 0m 3 / s.

[0019] 2) Operational boundary constraints

[0020]

[0021] Formulas (10)-(12) are the storage capacity, power generation flow, and output upper and lower limit constraints of power station m in period t respectively. V m and are the upper and lower limits of the storage capacity of power station m, respectively. 3 ; Q m and are the upper and lower limits of the power generation flow of power station m, respectively. 3 / s; and They are the upper and lower limits of the output of power station m, MW respectively.

[0022] 3) Storage capacity constraints from beginning to end

[0023] Given the initial and final storage capacity V of the cascade hydropower station during the dispatch period m,0 and

[0024]

[0025] In the formula, and are the storage capacity control values ​​at the beginning and end of the dispatch period set for power station m, 3 .

[0026] 4) Hydropower output function

[0027] In short-term scheduling, the daily water level fluctuation of the hydropower station has little impact on the water consumption rate. According to the actual operation of the hydropower station, the output is calculated using a fixed water consumption rate.

[0028]

[0029] In the formula, w m,t is the water consumption rate of hydropower station m in period t, m 3 / kW·h.

[0030] 5) Hydropower ramp constraints

[0031]

[0032] In the formula, Represents the ramp limit of hydropower station m, MW.

[0033] Step 4: Configure hydropower flexibility reserve to compensate for the forecast error of renewable energy output. In the context of limited probability distribution information (such as the mean and variance of random variables), the probability distribution of the forecast error of renewable energy output is unknown, but the mean and variance can be calculated based on sample data. Use moment fuzzy sets to model the forecast error of renewable energy output and make it satisfy the following distributed robust chance constraints:

[0034] 1) Hydropower upward regulation flexibility reserve constraints

[0035]

[0036] In the formula, ξ t represents the output forecast error of renewable energy in period t, MW; P represents the probability distribution of the output forecast error of renewable energy; P represents the fuzzy set where the potential probability distribution of the output forecast error of renewable energy lies; r t up is the hydropower reserve capacity during period t, MW; ε is the risk level at which the hydropower reserve cannot compensate for the forecast error of renewable energy output; Denotes the upward ramp limit of hydropower station m, MW.

[0037] 2) Hydropower down-regulation flexibility reserve constraints

[0038]

[0039] In the formula, r t down is the hydropower reserve capacity during period t, MW; Denotes the downward ramp limit of hydropower station m, MW.

[0040] 3) Moment fuzzy set of new energy output prediction error

[0041] P:={P:E[ξ t ]=μ,E[(ξ t -μ)(ξ t -μ)T]=Σ} (21)

[0042] Where E[] represents the expected value operator; μ and Σ represent the mean and variance of the forecast error of renewable energy output, respectively.

[0043] Step 5: Constraint Refactoring

[0044] The model in step 4 is an uncertainty optimization model, which needs to be converted into a deterministic model when solving. In addition, the relative prediction error is used in the model conversion process to ensure the generalization ability of the model.

[0045] 1) Relative prediction error of new energy

[0046]

[0047] Where P t RES Represents the predicted output of renewable energy in period t, MW; β t Represents the relative prediction error of new energy.

[0048] 2) Hydropower flexibility reserve constraint conversion

[0049] The conversion process of the hydropower upward flexibility reserve constraint is as follows:

[0050] As shown in equations (17) and (18), under the condition of compensating for the forecast error of renewable energy output, the reserve capacity of hydropower should be as small as possible. Since the ramping capacity of hydropower is a constant, situation, and take Combined with the relative prediction error, these two formulas are combined into the following formula:

[0051]

[0052] Next, we transform equation (23) into a deterministic constraint, where we introduce the Chebyshev inequality for transformation. The Chebyshev inequality is as follows:

[0053]

[0054] Where X represents an arbitrary random variable, whose mean and variance are E(X) and Var(X) respectively; r is an arbitrary non-negative number. It is called the Chebyshev inequality lower bound.

[0055] Combined with Chebyshev's inequality, equation (23) is transformed as follows:

[0056] First, let X = P t RES β t , then:

[0057] E(P t RES β t )=P t RES μ (25)

[0058]

[0059] Therefore, the right side of the equal sign in equation (23) can be changed to:

[0060]

[0061] In fact, the mathematical meaning of the above formula is the probability value corresponding to the worst scenario in the fuzzy set. According to the Chebyshev inequality, this probability value should be equal to the corresponding Chebyshev lower bound, that is:

[0062]

[0063] Among them, r satisfies the following conditions:

[0064]

[0065] Substitute equation (28) back into equation (23) and transform it to obtain:

[0066]

[0067] Therefore, the deterministic constraint form of the hydropower upward flexibility reserve constraint is as follows:

[0068]

[0069] The corresponding hydropower upward adjustment flexibility reserve capacity is:

[0070]

[0071] Similarly, the hydropower down-regulation flexibility reserve constraint can be converted to:

[0072]

[0073] The corresponding hydropower down-regulation flexibility reserve capacity is:

[0074]

[0075] Step 6: Steps 3 to 5 construct a short-term dispatch model for hydropower, wind and solar power, which is ultimately a deterministic linear programming model. By solving the model with an optimization solver, the output plan of cascade hydropower can be obtained, and the hydropower reserve capacity that compensates for the forecast errors of wind power and photovoltaic output can be quantified.

[0076] The results of the present invention have the following beneficial effects: when sample data is limited, the present invention models the uncertainty based on the limited probability distribution information (mean and variance) of the forecast error of renewable energy output; while performing peak load regulation, hydropower provides sufficient standby flexibility, promotes the consumption of renewable energy, and ensures the safe operation of the system. The present invention provides a short-term complementary dispatching plan for hydropower, wind and solar power, and ensures that the plan has good robustness and is still feasible in unknown scenarios, without increasing the applicability of the short-term dispatching plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 It is the overall framework diagram of the method of the present invention.

[0078] Figure 2 It is a schematic diagram of the peak regulation of water, wind and light during the dry season, among which (a) is the peak regulation diagram of the DRCCP model, (b) is the peak regulation diagram of the CCP_G model, and (c) is the peak regulation diagram of the CCP_SAA model.

[0079] Figure 3 It is a schematic diagram of the peak regulation of water, wind and light during the flood season, among which (a) is the peak regulation diagram of the DRCCP model, (b) is the peak regulation diagram of the CCP_G model, and (c) is the peak regulation diagram of the CCP_SAA model.

[0080] Figure 4 It is a schematic diagram of the output of each hydropower station during the dry season, among which (a) is the schematic diagram of the output of cascade hydropower stations in the LC basin, and (b) is the schematic diagram of the output of cascade hydropower stations in the JS basin.

[0081] Figure 5 It is a schematic diagram of the output of each hydropower station during the flood season, among which (a) is the schematic diagram of the output of cascade hydropower stations in the LC basin, and (b) is the schematic diagram of the output of cascade hydropower stations in the JS basin.

[0082] Figure 6It is a schematic diagram of the mean and standard deviation of the out-of-sample peak-shaving objective function value of each model at different confidence levels, where the broken line represents the change of the objective function value, and the shaded part around the broken line represents the standard deviation of the objective function value.

[0083] Figure 7 It is a schematic diagram of the mean load adjustment values ​​of each model at different confidence levels.

[0084] Figure 8 It is a schematic diagram of the load adjustment value distribution of each model at different confidence levels.

[0085] Fig. 9 It is a schematic diagram of the alternative constraint violation probability of each model at different confidence levels.

[0086] Fig.10 Schematic diagram of the hydropower output limit and flexible reserve capacity under different confidence levels of the method of the present invention, wherein (a) is the hydropower output limit and (b) is the hydropower flexible reserve. DETAILED DESCRIPTION

[0087] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0088] A short-term complementary scheduling method for hydropower, wind power and solar power under finite probability distribution information, the steps are as follows:

[0089] Step 1: Collect initial calculation conditions, including basic operating parameters of hydropower stations, coupled dispatching demand conditions of power system and cascade hydropower stations, forecast data of photovoltaic output, wind power output, load and interval water inflow at 96 points in the future day, and 96 points of measured historical data and forecast data of wind power output and photovoltaic output in the past three years.

[0090] Step 2: Analyze and organize the measured historical data and forecast data of wind power output and photovoltaic output in the past three years, and calculate the relative forecast error data of wind power output and photovoltaic output in the past three years. These relative forecast error data constitute the sample data set of uncertainty of new energy output, and the sample data set is divided into in-sample data and out-of-sample data in a ratio of 3:7. The output forecast error of new energy is the main source of uncertainty, and the flexibility demand of the power grid will be related to the forecast error of new energy output.

[0091] Step 3: Aggregate all wind farms and photovoltaic power stations connected to the grid into a virtual wind farm and a virtual photovoltaic power plant respectively. Utilize cascade hydropower to provide peak load flexibility and reserve flexibility, and couple cascade hydropower with wind and solar power to achieve short-term complementary scheduling. With the goal of minimizing the absolute value of the average deviation of residual load, a short-term complementary scheduling model for hydropower, wind and solar power is constructed:

[0092]

[0093]

[0094] Where: F is the objective function value, MW; C t , and are the original load, net load and residual load of the power grid in period t, MW; P t AW and P t AS are the predicted wind power output and photovoltaic power output in period t, MW; is the output of hydropower station m in period t, MW; C R,ave is the average value of residual load, MW; T and M are the total number of time periods and the total number of hydropower stations respectively. Introducing auxiliary variable δ t , perform equivalent transformation on formula (1):

[0095]

[0096] The constraints are as follows:

[0097] 1) Water balance constraints

[0098]

[0099] R m,t =q m,t +s m,t (9)

[0100] Where V m,t is the storage capacity of hydropower station m at the end of period t, m 3 ;I m,t and R m,t are the interval water flow and outflow flow of hydropower station m in period t, respectively. 3 / s; where the outbound flow includes the power generation flow q m,t and the abandoned water flow s m,t , m 3 / s; Δt represents the step length of the scheduling calculation period, which is 15 minutes; U m is the index set of the upstream power stations of power station m, j is the upstream power station number, R j,t is the outflow of upstream power station j in period t, m 3 / s. Since water abandonment is not allowed in hydropower stations, the abandonment flow rate is set to 0m 3 / s.

[0101] 2) Operational boundary constraints

[0102]

[0103] Formulas (10)-(12) are the storage capacity, power generation flow, and output upper and lower limit constraints of power station m in period t respectively. V m and are the upper and lower limits of the storage capacity of power station m, respectively. 3 ; Q m and are the upper and lower limits of the power generation flow of power station m, respectively. 3 / s; and They are the upper and lower limits of the output of power station m, MW respectively.

[0104] 3) Storage capacity constraints from beginning to end

[0105] Given the initial and final storage capacity V of the cascade hydropower station during the dispatch period m,0 and

[0106]

[0107] In the formula, and are the storage capacity control values ​​at the beginning and end of the dispatch period set for power station m, 3 .

[0108] 4) Hydropower output function

[0109] In short-term scheduling, the daily water level fluctuation of the hydropower station has little impact on the water consumption rate. According to the actual operation of the hydropower station, the output is calculated using a fixed water consumption rate.

[0110]

[0111] In the formula, w m,t is the water consumption rate of hydropower station m in period t, m 3 / kW·h.

[0112] 5) Hydropower ramp constraints

[0113]

[0114] In the formula, Represents the ramp limit of hydropower station m, MW.

[0115] Step 4. Due to the prediction error of renewable energy output, there will be a deviation between the short-term scheduling plan output of hydropower, wind power and solar power and the actual output. In order to avoid power imbalance caused by the prediction error of renewable energy output, hydropower flexible reserve is configured to compensate for the prediction error of renewable energy output. In the context of finite probability distribution information (such as the mean and variance of random variables), the probability distribution of the prediction error of renewable energy output is unknown, but the mean and variance can be calculated based on sample data. In order to solve the above problems, the prediction error of renewable energy output is modeled using moment fuzzy sets, and it satisfies the following distributed robust chance constraints:

[0116] 1) Hydropower upward regulation flexibility reserve constraints

[0117]

[0118] In the formula, ξ t represents the output forecast error of renewable energy in period t, MW; P represents the probability distribution of the output forecast error of renewable energy; P represents the fuzzy set where the potential probability distribution of the output forecast error of renewable energy lies; r t up is the hydropower reserve capacity during period t, MW; ε is the risk level that the hydropower reserve cannot compensate for the forecast error of renewable energy output, which is usually a smaller probability value, such as 0.05; Denotes the upward ramp limit of hydropower station m, MW.

[0119] 2) Hydropower down-regulation flexibility reserve constraints

[0120]

[0121]

[0122] In the formula, r t down is the hydropower reserve capacity during period t, MW; Denotes the downward ramp limit of hydropower station m, MW.

[0123] 3) Moment fuzzy set of new energy output prediction error

[0124] P:={P:E[ξ t ]=μ,E[(ξ t -μ)(ξ t -μ) T ]=Σ} (21)

[0125] Where E[·] represents the expected value operator; μ and Σ represent the mean and variance of the forecast error of renewable energy output, respectively.

[0126] Step 5: Constraint Refactoring

[0127] The model in step 4 is an uncertainty optimization model, which needs to be converted into a deterministic model when solving. In addition, the relative prediction error is used in the model conversion process to ensure the generalization ability of the model.

[0128] 1) Relative prediction error of new energy

[0129]

[0130] Where P t RES Represents the predicted output of renewable energy in period t, MW; β t Represents the relative prediction error of new energy.

[0131] 2) Hydropower flexibility reserve constraint conversion

[0132] The conversion process of the hydropower upward flexibility reserve constraint is as follows:

[0133] As shown in equations (17) and (18), under the condition of compensating for the forecast error of renewable energy output, the reserve capacity of hydropower should be as small as possible. Since the ramping capacity of hydropower is a constant, situation, and take Combined with the relative prediction error, these two formulas are combined into the following formula:

[0134]

[0135] Next, we transform equation (23) into a deterministic constraint, where we introduce the Chebyshev inequality for transformation. The Chebyshev inequality is as follows:

[0136]

[0137] Where X represents an arbitrary random variable, whose mean and variance are E(X) and Var(X) respectively; r is an arbitrary non-negative number. It is called the Chebyshev inequality lower bound.

[0138] Combined with Chebyshev's inequality, equation (23) is transformed as follows:

[0139] First, let X = P t RES β t , then:

[0140] E(P t RES β t )=P t RES μ (25)

[0141]

[0142] Therefore, the right side of the equal sign in equation (23) can be changed to:

[0143]

[0144] In fact, the mathematical meaning of the above formula is the probability value corresponding to the worst scenario in the fuzzy set. According to the Chebyshev inequality, this probability value should be equal to the corresponding Chebyshev lower bound, that is:

[0145]

[0146] Among them, r satisfies the following conditions:

[0147]

[0148] Substitute equation (28) back into equation (23) and transform it to obtain:

[0149]

[0150] Therefore, the deterministic constraint form of the hydropower upward flexibility reserve constraint is as follows:

[0151]

[0152] The corresponding hydropower upward adjustment flexibility reserve capacity is:

[0153]

[0154] Similarly, the hydropower down-regulation flexibility reserve constraint can be converted to:

[0155]

[0156] The corresponding hydropower down-regulation flexibility reserve capacity is:

[0157]

[0158] Step 6: Steps 3 to 5 construct a short-term dispatch model for hydropower, wind and solar power, which is ultimately a deterministic linear programming model. By solving the model with an optimization solver, the output plan of cascade hydropower can be obtained, and the hydropower reserve capacity that compensates for the forecast errors of wind power and photovoltaic output can be quantified.

[0159] The model and method proposed in this invention are verified by taking the hydropower, wind and solar power complementary system of a provincial power grid as the engineering background. All wind farms and photovoltaic power stations in the province are aggregated into wind power virtual power plants and photovoltaic virtual power plants respectively. The hydropower and wind and solar power of the province are used for short-term complementary scheduling. The hydropower stations are mainly distributed in two major river basins. The specific parameters are as follows:

[0160] Table 1 Hydropower station operation parameters

[0161]

[0162] In order to verify the robustness of the proposed model (DRCCP model) and ensure that the dispatch results can still be executed in unknown scenarios, the model is analyzed out of sample, and two comparative models CCP_SSA and CCP_G are set for comparative analysis. Among them, the CCP_SSA model uses the average approximate sampling method to solve the hydropower flexibility reserve opportunity constraint; while the CCP_G model assumes that the forecast error of renewable energy output follows a Gaussian distribution, thereby solving the hydropower flexibility reserve opportunity constraint. The simulation analysis results are as follows:

[0163] (1) According to the ratio of 0.3, the in-sample data set is randomly selected from the wind and solar sample data set for modeling. The typical water inflow process in the flood season and dry season is selected for calculation, and the corresponding in-sample water-wind-solar complementary peak regulation results and the output of each hydropower station are obtained, see Figures 2 to 5 . The peak-shaving indicators are used to analyze the effect of peak-shaving, including: maximum residual load, peak-to-valley difference of residual load, standard deviation of residual load, and load factor of residual load. The calculation results of peak-shaving indicators in flood season and dry season are shown in Tables 2 and 3. The following conclusions can be drawn from the in-sample calculation results: 1) When the confidence level increases, the peak-shaving performance of all models decreases, which means that the peak-shaving scheduling results become more conservative, thereby compensating the possible wind and solar output forecast errors with a higher guarantee rate. 2) The in-sample peak-shaving effect of DRCCP is not as good as that of CCP_G and CCP_SAA models, but this does not mean that the DRCCP model is invalid, and out-of-sample analysis is needed to illustrate the effectiveness of the model.

[0164] (2) Use the remaining wind and solar output sample data to perform out-of-sample analysis. The peak-shaving effect of the method of the present invention is evaluated based on four indicators, namely the average value of the objective function value, the standard deviation of the objective function value, the load adjustment value, and the probability of violation of the opportunity constraint. The out-of-sample analysis results are shown in Figures 6 to 9 . Figure 6 The out-of-sample objective function value mean and its standard deviation are shown; compared with the benchmark models CCP_G and CCP_SAA, the out-of-sample objective function value mean and standard deviation of the DRCCP model are smaller, which means that the out-of-sample peak regulation effect of the model is better and meets the optimization criteria. Figure 7 The load adjustment values ​​between the models were compared. Figure 8 The distribution of their load adjustment values ​​is statistically analyzed. It can be seen from the figure that under the same confidence level, the average value and distribution range of the load adjustment value of the method of the present invention are the smallest. Fig. 9 By comparing the chance constraint violation probabilities of various models, it can be seen that only the DRCCP model meets the reliability requirements.

[0165] Table 2 Peak regulation indexes of water, wind and solar power in dry season

[0166]

[0167]

[0168] Table 3 Water, wind and solar peak regulation indicators during flood season

[0169]

[0170] (3) Based on the hydropower flexibility reserve model in step 5, analyze the hydropower flexibility reserve at different confidence levels. Fig.10 As can be seen from the above, as the confidence level increases, the flexible reserve capacity of hydropower increases accordingly, thereby better coping with the uncertainty of wind and solar output. However, increasing the flexible reserve of hydropower will limit the output boundary of hydropower, thus affecting the peak-shaving performance. Therefore, it is necessary to set a reasonable confidence level to balance the flexibility of hydropower reserve and peak-shaving flexibility.

[0171] In summary, when the water, wind and solar power are complementary to each other in peak load regulation, the cascade hydropower stations not only provide peak load regulation flexibility to smooth the remaining load of the power grid, but also provide backup flexibility to deal with the forecast error of renewable energy output. The uncertainty of wind and solar power is modeled using limited probability information, and the corresponding distributed robust opportunity constraint model is constructed to solve the worst-case scenario, and finally obtain a short-term scheduling plan with better robustness.

[0172] The above-described embodiments merely express the implementation methods of the present invention, but they cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. A short-term complementary scheduling method for hydropower, wind power and solar power under limited probability distribution information. It is characterized in that Here are the steps: Step 1: Collect initial calculation conditions, including basic operating parameters of hydropower stations, coupled dispatching demand conditions of power system and cascade hydropower stations, forecast data of PV output, wind power output, load, and interval water supply at 96 points in the future day, and 96 points of measured historical data and forecast data of wind power output and PV output in the past three years; Step 2: Analyze and organize the measured historical data and forecast data of wind power output and photovoltaic output in the past three years, and calculate the relative forecast error data of wind power output and photovoltaic output in the three years; these relative forecast error data constitute a sample data set of uncertainty of new energy output, and the sample data set is divided into in-sample data and out-of-sample data according to the proportion; the output forecast error of new energy is the main source of uncertainty, and the flexibility demand of the power grid will be related to the forecast error of new energy output; Step 3: Aggregate all wind farms and photovoltaic power stations connected to the grid into a virtual wind power plant and a virtual photovoltaic power plant respectively; use cascade hydropower to provide peak-shaving flexibility and backup flexibility, and couple cascade hydropower with wind and solar power to achieve short-term complementary scheduling; and build a short-term complementary scheduling model for hydropower, wind and solar power with the goal of minimizing the absolute value of the average deviation of residual load: Where: F is the objective function value, MW; C t , and are the original load, net load and residual load of the power grid in period t, MW; P t AW and P t AS are the predicted wind power output and photovoltaic power output in period t, MW; is the output of hydropower station m in period t, MW; C R,ave is the average value of residual load, MW; T and M are the total number of time periods and the total number of hydropower stations respectively; the auxiliary variable δ is introduced t , perform equivalent transformation on formula (1): The constraints are as follows: 1) Water balance constraints R m,t =q m,t +s m,t (9) Where V m,t is the storage capacity of hydropower station m at the end of period t, m 3 ;I m,t and R m,t are the interval water flow and outflow flow of hydropower station m in period t, respectively. 3 / s; where the outbound flow includes the power generation flow q m,t and the abandoned water flow s m,t , m 3 / s; Δt represents the step length of the scheduling calculation period, which is 15 minutes; U m is the index set of the upstream power stations of power station m, j is the upstream power station number, R j,t is the outflow of upstream power station j in period t, m 3 / s; because the hydropower station is not allowed to abandon water, the abandoned water flow is set to 0m 3 / s; 2) Operational boundary constraints Formulas (10)-(12) are the storage capacity, power generation flow and output upper and lower limit constraints of power station m in period t respectively; V m and are the upper and lower limits of the storage capacity of power station m, respectively. 3 ; Q m and are the upper and lower limits of the power generation flow of power station m, respectively. 3 / s; and are the upper and lower limits of the output of power station m, MW; 3) Storage capacity constraints from beginning to end Given the initial and final storage capacity V of the cascade hydropower station during the dispatch period m,0 and In the formula, and are the storage capacity control values ​​at the beginning and end of the dispatch period set for power station m, 3 ; 4) Hydropower output function In short-term dispatching, the daily water level fluctuation of the hydropower station has little impact on the water consumption rate. According to the actual operation of the hydropower station, the fixed water consumption rate method is used to calculate the output; In the formula, w m,t is the water consumption rate of hydropower station m in period t, m 3 / kW·h; 5) Hydropower ramp constraints In the formula, represents the ramp limit of the hydropower station m, MW; Step 4: Configure hydropower flexibility reserve to compensate for the forecast error of renewable energy output. In the context of finite probability distribution information, the probability distribution of the forecast error of renewable energy output is unknown, but the mean and variance can be calculated based on sample data. Use moment fuzzy sets to model the forecast error of renewable energy output and make it satisfy the following distributed robust chance constraints: 1) Hydropower upward regulation flexibility reserve constraints In the formula, ξ t represents the output forecast error of renewable energy in period t, MW; P represents the probability distribution of the output forecast error of renewable energy; P represents the fuzzy set where the potential probability distribution of the output forecast error of renewable energy lies; r t up is the hydropower reserve capacity during period t, MW; ε is the risk level at which the hydropower reserve cannot compensate for the forecast error of renewable energy output; represents the upward ramp limit of hydropower station m, MW; 2) Hydropower down-regulation flexibility reserve constraints In the formula, r t down is the hydropower reserve capacity during period t, MW; represents the downward ramp limit of the hydropower station m, MW; 3) Moment fuzzy set of new energy output prediction error In the formula, E[·] represents the expected value operator; μ and Σ represent the mean and variance of the forecast error of renewable energy output, respectively; Step 5: Constraint Refactoring The model in step 4 is an uncertainty optimization model, which needs to be converted into a deterministic model when solving. In addition, the relative prediction error is used in the model conversion process to ensure the generalization ability of the model. 1) Relative prediction error of new energy Where P t RES Represents the predicted output of renewable energy in period t, MW; β t represents the relative prediction error of new energy; 2) Hydropower flexibility reserve constraint conversion The conversion process of the hydropower upward flexibility reserve constraint is as follows: As shown in equations (17) and (18), under the condition of compensating for the forecast error of renewable energy output, the reserve capacity of hydropower should be as small as possible; since the ramping capacity of hydropower is a constant, situation, and take Combined with the relative prediction error, these two formulas are combined into the following formula: Next, we transform equation (23) into a deterministic constraint. Here, we introduce the Chebyshev inequality for transformation. The Chebyshev inequality is as follows: Where X represents any random variable, whose mean and variance are E(X) and Var(X) respectively; r is any non-negative number; It is called Chebyshev's inequality lower bound; Combined with Chebyshev's inequality, equation (23) is transformed as follows: First, let X = P t RES β t , then: E(P t RES b t )=P t RES m (25) Var(P t RES b t )=P t RES2 S (26) Therefore, the right side of the equal sign in equation (23) can be changed to: In fact, the mathematical meaning of the above formula is the probability value corresponding to the worst scenario in the fuzzy set; according to the Chebyshev inequality, this probability value should be equal to the corresponding Chebyshev lower bound, that is: Among them, r satisfies the following conditions: Substitute equation (28) back into equation (23) and transform it to obtain: Therefore, the deterministic constraint form of the hydropower upward flexibility reserve constraint is as follows: The corresponding hydropower upward adjustment flexibility reserve capacity is: Similarly, the hydropower down-regulation flexibility reserve constraint can be converted to: The corresponding hydropower down-regulation flexibility reserve capacity is: Step 6. Steps 3 to 5 construct a short-term scheduling model for water, wind and solar power, which is ultimately a deterministic linear programming model. By solving the model using an optimization solver, the output plan of cascade hydropower can be obtained, and the hydropower reserve capacity that compensates for the forecast errors of wind power and photovoltaic output can be quantified.

Images